H-Sphere Troubleshooter

FAQ Revised: Tuesday 28 June 2005

Table of Contents

1. Control Panel
2. Web and FTP
3. WebShell
4. osCommerce
5. Webalizer
6. AWStats
7. phpBB
8. Miva
9. Mail
10. IP and Domain
11. User signup
12. SSL
13. Credit Card Processing
14. Quotas and Limits
15. Billing
16. Databases
17. Trouble Tickets
18. New Version Upgrades
19. Customization
20. Suspending and Resuming Accounts
21. Bandwidth or Traffic
22. Urchin
23. Windows

1. Control Panel

1.1. My H-Sphere was down, and when it finally came up, it gives me 500 error when accessed by browser.
Seemingly your PostgreSQL database is down. If you use Linux, restart it with the command:
etc/rc.d/init.d/postgresql restart FreeBSD users must use the following two commands: /usr/local/etc/rc.d/010pgsql.sh stop /usr/local/etc/rc.d/010pgsql.sh start

1.2. My system crashed, and when I try to restart the PostgreSQL database, it wouldn't come up.
Try restarting it, then run:
ps auxww|grep postmaster
If you don't see postmaster process running, PostgreSQL did not start up. Execute the following:
rm ~postgres/data/postmaster.pid
FreeBSD users must use the following two commands: rm ~pgsql/data/postmaster.pid
Restart postgres DB again.

1.3. I noticed that PostgreSQL is eating a lot of CPU resources.
Enter Posgress and run the VACUUM command. This will clean up your DB. Note that doing this will take your system down for about an hour, do it over the night or when you anticipate the smallest system load.

1.4. My SiteStudio doesn't work correctly.
These are the most typical reasons for problems with SiteStudio:
  • DISPLAY variable is not set
  • xvfrb is not running
  • Image Maker was not started
  • The information you enter for user to connect to his ftp account is invalid
  • The URL to access user's site is incorrectly entered
  • User's home dir is incorrectly entered
  • User's home dir has incorrect permissions
  • FTP and URL point to different sites
  • FTP password or login name, or server
  • FIrewall does not allow ftp connection
For some specific troubles, check with SiteStudio troubleshooter.

1.5. I have created a reseller account, logged in, but can't enter admin cp.
Make sure:
- you have configured DNS for resellers
; - you are using a registered domain.
If you need to use an unregistered domain for your own use only, add this domain to your hosts file (/etc/hosts on *nix systems).

1.6. My license count is off. It shows 11 and I am only using one.
Possibly, you have broken accounts in hsphere database in the "accounts" table that don't have the necessary respective records in the table "user_account". If this is the case:
  1. Log in your CP box as cpanel.
  2. Enter H-Sphere database:
    psql hsphere wwwuser
    with password for wwwuser.
  3. Run query:
    DELETE FROM accounts WHERE id NOT IN (SELECT account_id FROM user_account);
  4. Log in your CP box as root.
  5. Restart your CP.

1.7. What do I do if the /var/log/auth.log file each minute accumulates new errors such as:
Feb 27 08:32:17 localhost sshd[30614]: Illegal user admin from [CP's IP]
Feb 27 08:32:17 localhost sshd[30614]: Failed none for illegal user admin from [CP's IP] port 48170
Go to E.MANAGER -> P.Servers. Click the name of your physical CP server. The Login field should be empty.

1.8. javax.mail.AuthenticationFailedException: authorization failed. This keeps showing up in the hsphere.log file all the time.
If a reseller has incorrectly configured the support center (Support e-mail address, POP3 host, POP3 user, POP3 password), the following error will be written to hsphere.log:
2004-01-28 06:51:02,555 [TT Client: ] INFO psoft.hsphere.mail.MailClient - -->Get messages for reseller #846
2004-01-28 06:51:02,616 [TT Client: ] WARN psoft.hsphere.mail.MailClient - Trying to retrieve mail
javax.mail.AuthenticationFailedException: authorization failed

In this example, you would need to contact the reseller (account #846) and explain to him how to configure the support center.

1.9. My CP works slow in Internet Explorer, while it's OK in Mozilla/Netscape.
The latest IE 6 version has certain problems with caching images. Try to set cache settings to "Automatically", and also increase memory for file caching.

1.10. Reseller has accidentaly deleted his Admin account
  1. Log into the system Postgres.
  2. Run the following queries:
    SELECT id FROM users WHERE username='<UnfortunateReseller>';
    SELECT admin_id FROM resellers WHERE id=<id>;
    DELETE FROM users WHERE id=<admin_id>;
    UPDATE resellers SET admin_id=0 WHERE id=<id>;
    UPDATE signup_guard SET flags=0 WHERE reseller_id=<id>;
  3. Restart control panel

2. Web and FTP

2.1. I am having problems with the web server.
Restart apache by running:
The script will try to fix all posible problems and start apache.

2.2. My web server crashed and does not restart with apache-reconfig.
Run killall -9 httpd, and execute apache-reconfig again. If this doesn't help, go to the log files directory, and see if it contains files over 2GB in size, and remove them if there are any. Usually error_log, su_exec & ssl_engine_log would be like that. Then execute apache-reconfig again. Versions 2.1 and above will rotate these files automatically.

2.3. Our Linux Machines have stopped allowing FTP access.
Try restarting Web Server and FTP as suggested in http://www.psoft.net/HSdocumentation/sysadmin/understanding_web_server_configuration.html.

2.4. User dedicated FTP doesn't work (login failed)
The /etc/pam.d/ftp is not configured properly. You need to change the entry for this file to:
auth     required  /lib/security/pam_listfile.so item=user sense=deny file=/etc/ftpusers onerr=succeed
auth     required  /lib/security/pam_pwdb.so shadow nullok
account  required  /lib/security/pam_pwdb.so
session  required  /lib/security/pam_pwdb.so
If you are a FreeBSD user, FreeBSD does not support PAM session directives. If you remove the following line from the FreeBSD section of README.PAM, PAM should work properly under recent versions of FreeBSD:
ftp session required     pam_unix.so     try_first_pass
Look here for more details on the issue.

2.5. My customers are having problems with FrontPage.
Try these tips from our customers:
- If only one website is affected, back it up, then try the following:
  1. Make sure all files and directories have correct ownership and permissions, next log into user's CP -> web Options and run the FrontPage Fix.
  2. If this doesn't help, go to the website directory and remove the directories that start with _vti_, next turn FrontPage off and back on and re-publish the website.
- If the problem affects multiple FP websites, try turning off all RLimit options in the plan editor. Next, login as each user and click Apply Changes on the Web Options page to repost configs.

2.6. After the upgrage to FrontPage 2002, my customers have their web pages displayed incorrectly, appearing to duplicate data throughout the page.
This is a bug in FrontPage 2002, which can be removed with the latest service pack for Microsoft Office XP. For additional information, please see the following article in the Microsoft Knowledge Base:
Q307841 OFFXP: How to Obtain the Latest Office XP Service Pack

2.7. When publishing sites with FrontPage, my customer is getting the error "cannot open we80.cnf no such file or folder"
First, switch FrontPage OFF for this customer. Then find and delete all frontpage directories and .htaccess files:
        find ./ -type f -name ".htaccess" -exec rm -f {} \;
        find ./ -type d -name "_vti*" -exec rm -rf {} \;
Then switch FrontPage ON.

2.8. I have a disk quota failure.
On the partition where quota is enabled, execute:
quotaoff -a
quotacheck -am
quotaon -a
After that, reset quota (or if all you want to do is to reset quota) on the control panel server by executing as root:
# su - cpanel -c "java psoft.hsphere.tools.SetQuota -lid N" where "N" - Logical Server ID where quota should be set
This will reset quota for all users on the system on all servers, so it might take quite a while to execute.

2.9. Our Cold Fusion server will not relay email for one of our clients that had us point the MX record to their Exchange server.
Try disabling mail service in the client control panel before the redirct to another mail server will work. If it doesn't help, make sure that the file ~vpopmail/etc/tcp.smtp has the line for the IP of your ColdFusion server (make sure RELAYCLIENT is in the upper case, and has no spaces). After that, in the directory ~vpopmail/etc/ execute:
/hsphere/shared/bin/tcprules tcp.smtp.cdb tcp.smtp.tmp < tcp.smtp
Read more here.

2.10. My customer can't run .php and .html files inside /cgi-bin.
You can't run HTML and PHP files from the /cgi-bin directory. This directory is only for executing CGI scripts.

2.11. I have a customer who is having problems with her perl scripts.
Make sure the user has the .pl extension added on the Web Options page in the user Control Panel. If this doesn't help, check if the path to perl is correct, the script files have executable permissions, and script code has Unix-style end of lines.

2.12. I get "java.lang.Exception: User is not an owner of [directory/file]
This error returns when a user's directory has a file or a subdirectory with a different ownership, and this subdirectory or file is being deleted by this user. The solution is to change ownership on this file or subdirectory.

2.13. FTP starts with the error: no such user 'username' or unknown user 'username'
It may happen sometimes when a user was deleted, but its virtual FTP config was not.

To fix this problem:

  1. Log in as cpanel:
    su -l cpanel
  2. Run the Java tool to repost FTP configs:
    - either by the problematic user's account id:
    java psoft.hsphere.tools.PostFTPConfigs -acc <account_id>
    - or by the logical server id:
    java psoft.hsphere.tools.PostFTPConfigs -lid <lserver_id>
    - or for all virtual FTPs:
    java psoft.hsphere.tools.PostFTPConfigs -all
  3. Restart FTP.

3. WebShell

3.1. When I'm trying to load WebShell, my browser gives an error.
WebShell uses Java scripts from the hosting server. If these scripts are inaccessible, the browser will give this error. The reason could be, for example, incorrect server configuration.

3.2. WebShell times out when users launch it.
  1. Ping your web server to make sure it's up and running.
  2. Check if apache/IIS is running on it (port 80). If it is, check with the logs.
  3. Make sure that your FTP listens to all IPs.
  4. Make sure your PHP is version 3.0 or later.
  5. On Windows servers, make sure your PHP was installed from zip archive, not exe binary.

4. osCommerce

4.1. I'm having problems with osCommerce.
Try enabling globals in your /usr/local/lib/php.ini by replacing register_globals = Off with register_globals = On and then restarting your web server.

4.2. I got this error: psoft.hsphere.resource.mysql.UserPrivileges.revokeAllDatabasePrivileges(UserPrivileges.java:140)
Please ignore this message. It doesn't mean you've got a problem on your server.

5. Webalizer

5.1. Webalizer stopped updating.
You can locate the problem by doing the following:
  1. Using WebShell or any other FTP manager, check the Last Modified date.
  2. If you see that the file hasn't been modified during the problem period, check that all files in the directory /hsphere/shared/apache/conf have entries for the problem domain.
  3. If the entries are missing, the customer needs to turn Webalizer off and back on.

5.2. Webalizer doesn't work for certain users.
One possible reason could be that the user has reached the disk quota or there is no disk space left on the server.

5.3. Webalizer statistics doesn't update for certain users.
This might be due to a frequent Windows server reboot. Webalizer processing log files for this user is interrupted and when it is lauched next time, it starts processing the same log files for that same period.

5.4. My client has enabled webalizer, but when he tries to access it through http://domain.com/webalizer/ he gets the error: "Forbidden, You don't have permission to access /webalizer/ on this server."
This can happen if you set a directory index in the root folder to something other than html, shtml etc. You can either put the index.html on the end of the url, or put a htaccess file in the webalizer folder setting index.html as the directory index.

5.5. Webalizer does not resolve the country
Step 1.
Change run_webalizer function in the cron_rotate.pl script (under root):
# vi /hsphere/shared/scripts/cron/cron_rotate.pl
and find "sub run_webalizer {". Add the following command to the second line after the function declaration:
&&safe;_suexec("root","/hsphere/shared/bin/webalizer -p -Q -N
20 -D/var/log/dns_cache.file -n $site -o $dest $file");
It would look as follows:
sub run_webalizer {
  &&safe;_suexec("root","/hsphere/shared/bin/webalizer -p -Q -N 20
  -D/var/log/dns_cache.file -n $site -o $dest $file");
Step 2.
Update the hsphere-webailizer package to the latest version according to the operating system:
wget http://www.psoft.net/shiv/HS/RH72/hsphere-webalizer-2.01-10.rpm
rpm -ivh --force hsphere-webalizer-2.01-10.rpm Step 3.
Run cron_analyze.pl and cron_rotate.pl scripts (this step could be skipped, as these scripts run each day as cron jobs):
# /hsphere/shared/scripts/cron/cron_analyze.pl
# /hsphere/shared/scripts/cron/cron_rotate.pl

6. AWStats

6.1. Problems with AWStats
Somethimes problems with AWStats are caused by enabled RLimitMEM. Try disabling this resource in the plan if it is enabled.

7. phpBB

7.1. Error creating new session in PhpBB.
This is a known bug in PhpBB2. The php_sessions table in the PhpBB MySQL database needs to be cleared regularily. Do this either manually using PhpMyAdmin tool, or create the corresponding script and add it to cron. For more details, refer to the PhpBB forum.

8. Miva

8.1. I successfully installed Miva, but it still doesn't work.
1. Disable RLimitCPU, RLimitMEM, RLimitNPROC for the plan that allows MIVA.
2. Republish Apache config file on the web server that Miva is installed on:
su - java
java psoft.hsphere.tools.PostApacheConfigs -lid X
where X is the ID of the server that Miva is installed on.

3. Restart Apache on the web server Miva is installed on:
For Linux:

/etc/rc.d/init.d/httpd stop
wait a while
/etc/rc.d/init.d/httpd start

For FreeBSD:

/usr/local/etc/rc.d/apache.sh stop
wait a while
/usr/local/etc/rc.d/apache.sh start

8.2. When a user enables Miva Merchant, the system returns "Miva Merchant isn't properly configured".
Make sure that /hsphere/shared/miva/ contains the Merchant-bundle link pointing to /hsphere/shared/miva/Merchant-v4.12.tar.gz
If it is missing, log into the web server that Miva Merchant is installed on as root, copy file Merchant-v4.16-bundle.tar.gz to /hsphere/shared/miva and execute:
ln -s /hsphere/shared/miva/Merchant-v4.16-bundle.tar.gz Merchant-bundle

8.3. Miva license counter doesn't work.
You probably have H-Sphere installed with NAT. Unfortunately, in this case H-Sphere is not able to trace the moment when a license is added. So, your counter would show 0, though your licenses are added to the system successfully.

9. Mail

9.1. My users can't send e-mails, and get the error "553 sorry, that domain isn't in my list of allowed rcpthosts"
To fix it, do the following:
  1. log into your mail server as root;
  2. stop qmaild:
    # /etc/rc.d/init.d/qmaild stop
  3. backup all needed files in the ~vpopmail/etc directory;
  4. delete the directory ~vpopmail/etc:
    # rm -rf ~vpopmail/etc
  5. create it anew:
    # mkdir ~vpopmail/etc
  6. copy backed files to this new directory;
  7. empty the file ~vpopmail/etc/open-smtp:
    # cat /dev/null > ~vpopmail/etc/open-smtp
  8. start qmaild:
    # /etc/rc.d/init.d/qmaild start

9.2. Our H-Sphere sends letters to customers ending like this:
"If you have troubles using our product, you can communicate them to us at the following details:"
How do I get it to show my company info?
You have to specify it in your admin Control Panel -> Settings menu -> Company info.

9.3. We are having problems with people using Outlook/Express. They are receiving a rcpthosts error.
Usually, we install H-Sphere Mail server with POP before SMTP feature included. This means that users need to check their pop3 box before sending mail trough the server. When the POP3 box is checked, user can send emails within the next 15-30 minutes.

9.4. My customers can't access their POP3 accounts with their e-mail clients. They get this error: "There was a problem logging onto your mail server. Your username or password was rejected. Server response: "ERR aack, child crash".
Your open-smtp file got corrupted. In SMTP-after-POP, this file is responsible for opening SMTP relay for specific IP addresses. When user conects by POP3, his IP address is saved in that file. Once the user tries to send anything, this file is checked for the users IP address.

To fix it, execute the following commands:
cd /hsphere/local/var/vpopmail/etc
cat /dev/null > open-smtp
/hsphere/shared/bin/tcprules tcp.smtp.cdb tcp.smtp.tmp < tcp.smtp
chown vpopmail:vchkpw *

9.5. vmoduser is running in the system, eating 70% of CPU, and causing mail server to be exteremely slow.
Kill vmoduser, make sure that the directory ~vpopmail/users/ doesn't have any files excluding .vpasswd.lock. You might remove .vpasswd.lock as well.

9.6. The mail server auth daemon stops accepting connections.
SqWebMail doesn't always integrate with RedHat 7.3 smoothly, so we recommend using IMP instead.

9.7. A user can't add a maibox because this mailbox reportedly exists, but it's not in the control panel.
If you don't see a mailbox in the CP but it exists on the mail server you can do the following:
1. Log in your mail server as root.
2. Change directory to ~vpopmail/bin/
3. Check if this mailbox exists:
# ./vuserinfo box@domain.com
4. If it exists and you don't see it in the Control Panel interface, you can try to delete it:
# ./vdeluser box@domain.com
5. Try to create this mailbox through H-Sphere Control Panel.

9.8. When trying to launch IMP, users get: "Login failed for some reason. Most likely your username or password was entered incorrectly."
It's a known IMP problem. Users have to click the "Logout" icon (not close window) every time when they finish working with IMP. To get around the problem, users should insert login, password, and server IP or try clicking the LAUNCH icon again.

9.9. Customers can't connect to their mailboxes using Outlook and there quotas are showing 0
Enter the maildomain directory:
cd ~vpopmail/domains/domain.com
and check the vpasswd file. It must be in plain text format, each line representing a record for a domain. If it's corrupted, do the following:
  1. remove broken vpasswd, .vpasswd.lock and vpassvd.cdb files;
  2. Connect to H-Sphere database and check passwords for clients mailboxes using the following query:
    select * from mailboxes where full_email like '%domain.com';
  3. Backup the maildomain directory:
    cd ~vpopmail/domains
    cp -r domain.com domain.com.back
  4. Remove the maildomain:
    ~vpopmail/bin/vdeldomain domain.com
  5. Add the maildomain:
    ~vpopmail/bin/vadddomain domain.com [postmaster's password]
  6. Add clients mailboxes listed in the DB query:
    ~vpopmail/bin/vadduser mailbox_name@domain.com [password]
  7. Remove the mailbox directories for just created mailboxes:
    rm -rf domain.com/mailbox_name
  8. Restore the mailbox directories from the backup:
    mv domain.com.back/mailbox_name domain.com/
  9. Move .qmail* files if needed
  10. Fix the maildirectory permissions:
    chown -R vpopmail:vchkpw domain.com
  11. Remove the backup directory:
    rm -rf domain.com.back
  12. Check the results:
    ~vpopmail/bin/vuserinfo mailbox_name@domain.com

9.10. Mail is not delivered from control panel
Step 1. Check SMTP settings:
cat ~cpanel/shiva/psoft_config/hsphere.properties | grep SMTP
SMTP_HOST = mail.domain.com
bash-2.05a$ telnet mail.domain.com 25

If your SMTP is configured correctly, this will produce a result similar to this:
Trying [IP address]...
Connected to mail.domain.com.
Escape character is '^]'.
220 mail.domain.com ESMTP

Step 2. If you have successfully connected with telnet, check if all mail settings in H-Sphere Control panel are configured properly. Step 3. If your CP mail settings are configured properly, try to create a test trouble ticket, answer it and check /var/log/hsphere/hsphere.log to see if it contains a record with "Sending email to: [e-mail address]host:
mail.domian.com from: [H-Sphere admin e-mail address]"

Step 4. If this record is there, check if the mail server allows relays from the CP server IP (E.Manager)

9.11. User cannot send mail from a webserver
If mail and web are installed on the same box, mail is serviced by qmail.
Step 1. Check /var/log/maillog for errors referring to the sender
Step 2. Log into the user CP, go to Web Options page for the domain mail is sent from, and view Error Log for errors related to sending mail.
The most common error is that there's no mail relay from this mail server to the web server IP.
If mail and web are installed on separate boxes, mail is serviced by SendMail, not qmail. Check that the file /etc/sendmail.cf has the directive DS[MAIL_SERVER_NAME], for example DSmail.example.com. If it doesn't, add this directive manually. This file can also be located in /etc/mail/

9.12. User is having problems with Frontpage mail
Frontpage problems are most frequently caused by RLIMITs and can be resolved by turning them off in plan settings. If end users report they can't send mail through Frontpage forms, most probably Frontpage Extensions are configured without the following parameters: SMTPHost, MailSender, MailReplyTo, SendMailCommand. This can be fixed with the FrontPage Extensions patch:

9.13. Support center hits an autoresponder and fills up with new tickets
Log into your CP as main admin, go to Settings -> TechSupport and check "Don`t send confirmation emails for received Trouble Tickets". In this case no messages will be sent to confirm that support request has been received. If the Support Center got overloaded with new tickets, retrieve them with the search utility and apply mass closure/deletion as instructed in www.psoft.net/HSdocumentation/admin/providing_technical_support.html#mass_action

9.14. Fighting spam
Instructions at: http://www.psoft.net/HSdocumentation/sysadmin/fighting_spam.html
Spamguard doesn't come with standard H-Sphere installations and can be installed optionally. Spamguard sources are downloaded on each mail server the customer wants to protect: http://www.enderunix.org/spamguard/ and is installed as instructed on this site according to customer's requirements. More at:

9.15. Large Mail Queue
To check the number of e-mails in queue, execute:
If the number exceeds, say, 1000, then check if your mailserver isn't attacked by spammers (see above).

9.16. Sending a letter generates several copies
This may happen when qmail server is overloaded. Try to reduce the server load.

9.17. Sending mail is delayed
This may happen when there is no reverse DNS record for the mail server main IP. Set reverse DNS record in the reverse zone file.

10. IP and Domain

10.1. IPs on my H-Sphere servers are down.
Information about H-Sphere IPs in use is stored in the /hsphere/local/network/ips file. If an IP goes down, run the setup-ips.pl script from the /hsphere/shared/scripts. After the server reboot, IPs are automatically put up by this srcipt. If it doesn't work, make sure this script is added to the /etc/rc.local file.

10.2. When I add IP's to the logical web server, the IP's do not get registered in the ifconfig.
Ips are added to Ipconfig only when a new account is created, not when they are assigned to a web server.

10.3. The user signup notification e-mail calls ns2 the primary nameserver and ns1 the secondary nameserver.
The distribution of primary/secondary DNS servers is done randomly to reduce load on particular DNS servers. As a result, a part of your customers will have ns2 the primary DNS server.

10.4. I run FreeBSD, and my dedicated IPs are inacessible.
Make sure your netmask is for dedicated IPs.

10.5. A user who has just signed up has multiple IPs.
This is how we think the problem appears:
1) A user creates a domain (domain.com) with the default shared IP. Two DNS records are created in the DNS zone, one for domain.com, the other for *.domain.com.
2) The user receives a notification e-mail which reads the location of the index.html page (domain.com).
3) The user goes to this page to see how it works. During this process the DNS query sends through DNS server chain to your server and the response is cached on each server of this chain. Now the DNS servers remember the location of domain.com at this shared IP.
4) The user goes to the control panel and changes to the dedicated IP.
5) The user goes to www.domain.com (which is not domain.com) and this domain name is mapped to the dedicated IP. But domain.com still points at the shared IP, and it will be so until the DNS servers update themselves, which takes at least 24 hours.

The solution is to wait until the DNS servers update themselves. To avoid this problem, the user has to do steps 1, 3, 4. Step 2 must be skipped.

10.6. I am trying to add a domain zone into H-Sphere but it returns the "update failed" error.
Are you sure that H-Sphere DNS server doesn't already contain this zone? Check the named.conf named configuration file. If it contains this zone, remove it and try to add it in the H-Sphere CP again. If it doesn't help, see the next question.

10.7. I am trying to add a service zone, but get a NullPointerException error.
Make sure all your name servers have Service IPs.

10.8. Nobody can use my nameservers.
If you have a firewall, open port 53 udp packets. Also, open port 53 for tcp and udp packets for localhost and slave nameservers, otherwise H-Sphere and slave nameservers won't be able to work properly.

10.9. My DNS server worked fine, but H-Sphere started returning errors several hours ago.
Do you have the named daemon started? Do you have the bin8.x version installed or you use the RedHat autoupdate which installed the bind9.x version instead of the old one?
H-Sphere doesn't support bind9.x versions!!!
Back up the named.conf file, uninstall the bind9.x package, install the bind8.x package instead of it and place the named.conf file back; restart the named daemon after it.

10.10. Whenever I reboot hsphere server, all IP based accounts would stop working.
Run /home/hsphere/shared/scripts/setup-ips.pl

10.11. When I register a domain through Enom, I get "Failed to add Account, Invalid Domain Name or Nameserver".
There can be several reasons:
- Your nameserver isn't registered with InterNIC. Make sure it's registered and the propagation period has elapsed. - Your registered nameservers don't match those entered in the control panel. For instance, your registered nameservers are ns1.example.com and ns2.example.com, and in your control panel, you have configured ns.example.com and ns2.example.com - The domain you are trying to register is already taken;
- You haven't installed the Enom patch from http://www.psoft.net/misc/enom.html.

10.12. When I register a domain, I get this from the registrar server: "Invalid data Missing required field: billing first_name Missing required field: billing last_name..."
Make sure you have filled in 'Owner first name' and 'Owner last name' on the Look and Feel -> Company Info page.

10.13. Reseller control panel is unavailable by reseller's instant alias
Make sure that that base domain has the same nameserver both on H-Sphere and with the registrar (run whois domainname.com)

10.14. We have over 500 DNS IPs, and the named returns internal server errors. Users can't add dns records or properly create domains.
It's a known problem and the solution is to use both Dedicated and Shared IPs for reseller DNS. If a resellers' domain registrars don't require dedicated DNS IPs, you can have the reseller use shared DNS IPs. The details are given at http://www.psoft.net/HSdocumentation/admin/disabling_resources.html#dns_ip To move a reseller to a shared IP, you need to change the plan configuration and recreate reseller's server aliases for his DNS boxes manually. If you do it, don't forget to ask the reseller to change his domain registrar DNS data accordingly.

10.15. Users can't switch to dedicated IPs.
Make sure that you have free dedicated IPs available on this server (you can check it on the E.Manager page)
The user's plan may restrict the number of dedicated IPs per account or the Dedicated IP resource is disabled in the plan. To check it, go to INFO->Plans->Resources Make sure that Dedicated IP is listed in enabled resources. If it isn't, enable it. If it is enabled, click the Edit link next to it. Max value for dedicated IPs should be more than the customer is currently using.

10.16. User registering domains through H-Spehreare get either of the following errors:
ERROR!!! Response from registrar server: Not supported on this cipher.:-2008
ERROR!!! Response from registrar server: Unsupported keysize or algorithm parameters:-2008
If you are running Java v 1.4.2, make sure you have downloaded "Unlimited Strength" Jurisdiction Policy Files from http://java.sun.com/j2se/1.4.2/download.html#docs, section "Other Downloads".
and put them in the directory JAVA_HOME/jre/lib/security
where JAVA_HOME is the Java SDK home directory.

10.17. Name server check on domain registration returns a DNS error that reads "Domain has no MX or A record".
Make sure logical mail server (or reseller mail server alias) is registered in the DNS zone managing it as an A record or MX pointing to an A record, not CNAME.
In H-Sphere 2.3 RC 4 or higher resellers can do it by simply recreating their mail server aliases.

10.18. Upon changing service domain (service DNS zone), instant aliases for new and existing domains are created based on an old one's instant alias template.
Go through the plan edit wizards to repost plans' conifgs and then using mass mail ask existing users to turn on and then back on instant alias on Web Options page for each of their domains.

11. User signup

11.1. What do I do if at the signup or when adding a certain resource I get "No host available for signup" error message?
First, make sure the server is present in the list of servers. If it is, make sure to confirm the changes in the system configuration by clicking the "Apply changes" link both for physical and logical servers. Also, make sure the "Available for signup" button is ON in the logical server setting forms . This option is enabled by default for new logical servers.

11.2. We are unable to sign up new accounts. We are receiving a "Generic Error". What should we do to fix this problem?
Try restarting the control panel.

11.3. I am trying to set up a Real Server Unix account. While trying to moderate it for the first time, I get the error message "Account could not be added: There is no host available for signup".
You have probably forgotten to apply changes to the server configuration. This is done by clicking the Apply Changes link under the logical server list on the E.Manager page.

11.4. I sign up an account, then go into CP as Admin and can't find it there through Search.
If Account moderation is enabled, go to Signup->Moderated and modetate the account.

11.5. When I try to create account I get "Duplicate login! The user name than you entered already used. Please, enter new user name" but I can't find any duplicates.
It could have been signed up as moderated and wasn't activated, that's why you can't find it. In that case you'll find it in Signup -> Moderate. If you can't find it in the "awaiting moderation" list, click here.

11.6. I was signing up an account and got an error: "Failed to create Account. The previous request for the user creation has been locked in. If it failed, try to sign up from scratch."
If a signup fails, it is added to the list of Accounts to be moderated (the Signup section). In this case, all you need to do is delete this account in the list and re-try the signup. If the account is not in the list, simply close the signup wizard and re-try the signup from scratch.

12. SSL

12.1. I would like to use ComodoSSL, but I'm getting "com.comodo.api.ComodoException: The value of the 'loginName' argument is invalid!"
Make sure you have registered with Comodo as a "Reseller" and not as a "Web Host Reseller".

12.2. My user enabled Shared SSL for his domain and created 3rd Level Domain Alias, however when he goes to the page by https it fails to be displayed saying it "can't find server".
Make sure the Domain alias A DNS Record resource isn't disabled for this plan, as it is essential for 3rd Level Domian Alias which in its turn is required to enable Shared SSL. DNS zone for this 3rd Level Domain Alias should contain two built-in records pointing to the IP address of the main domain.

12.3. When a user goes to his site through https he gets the following Security Information message "This page contains both secure and nonsecure items. Do you want to display nonsecure items?"
Make sure the page source doesn't contain any kinds of http links.

12.4. My user is trying to install a certificate and key on the site and gets the following error: Invalid certificate format, file: server.key
To install your certificate and key on your site, you need to remove encryption and enter it in clear-text format. Encryption can be removed with the OpenSSL utility of Unix Shell on any Unix server where OpenSSL utilities pack is installed, for example on the Control Panel box.
  1. Put you encrypted Private Key into a separate file and save it as server_old.key
  2. In Unix Shell, execute the following command: openssl rsa -in server_old.key -out server.key
  3. Enter the password when prompted.
  4. A new file (server.key) will be generated. It will contain the Private Key in a plain text format.

12.5. Installing a certificate authority file produces an error.
You may be getting the following error:Invalid certificate format, file: ca-bundle.crt
This has been fixed in H-Sphere 2.3.1 beta 2. When adding the reseller certificate to the system, a line was missing in the reseller SSL configuration file. To fix this problem immediately, add the following line to /hsphere/local/home/cpanel/conf/sites/[resellername].conf:
SSLCertificateChainFile /hsphere/local/home/cpanel/conf/sites/[resellername]/ca.crt
Then restart H-Sphere.

13. Credit Card Processing

13.1. I use Authorize.net. Why are my customers getting a "First Name Cannot be Left Blank" error, even though they are typing in their first name?
Please refer to http://www.authorizenet.com/support/vsr/5/2/3/

13.2. I use Authorize.net. Why am I getting the error: "Merchant account not present or invalid"?
You cannot perform two-phase transactions in the test mode. You should either ignore these messages or go live.

13.3. The script which connects to Authorize.net is doing a standard charge, then attempts to charge the card from the previous authorization.
Use comma to delimit the results from Authorize.net.

13.4. I am using PayAndShop. Why am I getting the error: "PostAuth(CC Capture) error after successful third party resource (like OpenSRS) registration"?
Try to contact PayAndShop, and ask them to anable "settle" functionality for your account.

13.5. When the system tries to charge my customer's credit card, it returns: "Error processing credit card: Expected entity name for reference (line 1, column 406)"
When registering, your customer entered an invalid Credit Card Name. It can have only first name and last name separated with a whitespace. It can't have a middle or a title. For example, the name George W. Bush is wrong and must be changed to George Bush.

14. Quotas and Limits

14.1. We have an account where their bandwidth was limited to 5GB. However, they were able to exceed their limit.
In most resources, such as disk space, limit cannot be exceeded. However, this is not the case with the traffic limit. It can only be charged at a higher rate - the usage fee.

14.2. I Just changed disk quota in one of our plans and our customers were not affected by this change.
If you want these changes to affect some of your existing users, you need to enter their control panels one by one, go to Web Options page and click Apply changes.

14.3. The control panel says: "Disk quota Used 110 MB of 120 MB limit", but the actually used disk space is only 55 MB.
H-Sphere calculates disk quota usage based on file ownership, not user's home directory content. When an account is deleted, its UID is can be assigned to a new user. This new user will become the owner of the old user's files if they weren't deleted with the account. To find out the UID of this user, execute the following commands:
# id unixtest
You will get something similar to the following:
uid=1975(username) gid=1972(username) groups=1972(username)
Then you can get the list of files that belong to this UID:
# find /hsphere/local/home -uid 1975 -print
You can remove the files that are located outside this user's home directory or chown them to root.

14.4. One of my customers shows disk quota Used 0 MB of 0 MB limit, but actually has about 5MB.
The disk space problem appears when a check-paying user tries to increase the disk space quota. Disk quota resource is changed on the add/remove basis. The old disk space quota is removed and the new disk space quota fails to be created but fails, because this would go beyond the credit limit. The solution is to increase this user's credit limit and then restore the disk quota resource.

14.5. One of my customers shows total traffic Used 0 KB of 6.0 GB limit, but actually has used up to 100MB.
Traffic is calculated and reset on monthly basis, regardless of the billing period. Obviously, this customer's traffic was recently reset.

14.6. My user has 0 MB used out of -1 MB in unlimited files.
This problem is caused by broken web server disk quota. To fix this:
1) Log into the web server as root:
su -
2) Turn off disk quota:
quotaoff [hsphere_partition]
3) Go to H-Sphere partition and remove files aquota.user and quota.user:
cd [hsphere_partition]
rm -f aquota.user quota.user

4) Update aquota.user and quota.user files::
quotacheck -m [hsphere_partition]
  * ignore "-m" for Free BSD
5) Turn on disk quota:
quotaon [hsphere_partition]
6) Then log into CP server as root and change to cpanel user:
su - l cpanel
7)Reset user defined quotas:
java psoft.hsphere.tools.SetQuota -lid [logical_server_id]

If your H-Sphere is installed on the root partition, the commands will look as follows:
quotaoff /
quotacheck -m /

  * ignore "-m" for Free BSD
quotaon /

15. Billing

15.1. My customers are not billed. What's the problem?
One possible reason can be that your customers have not reached their credit limit. Go to this plan's edit wizard and see the value of this parameter. Customers will be charged authomatically when their negative balance reaches the credit limit.

15.2. Account event was supposed to call the transaction, but it didn't.
Probably, invoicing hasn't been performed yet. To check when accounting cron was last launched:
  1. Log into your CP box as cpanel user:
    su -l cpanel
  2. Log into your PGSQL "hsphere" database as wwwuser:
    psql hsphere wwwuser
  3. Execute:
    select LAST_LAUNCH from last_start;
    This will show the accounting cron and the time it was last run.
If it was run before the account event, you'll just have to wait until it is launched next time.

15.3. Free units and prices in Reseller Price, reseller top-level account are different from those in the reseller plan edit wizard.
Probably, reseller plan ID got improperly updated on a certain account event. To check if the plan_id in the "resellers" table is the same as in the "accounts" table:
  1. Log into your CP box as cpanel user:
    su -l cpanel
  2. Log into the "hspere" system database as wwwuser:
    psql hsphere wwwuser
  3. Find this reseller ID in the "users" table:
    select id from users where username='X';
    where X is this reseller's login.
  4. Get the reseller plan ID from the "resellers" table:
    select reseller_plan_id from resellers where id='X'
    where X is this reseller's ID retrieved on step 3.
  5. Get the reseller plan ID from the "accounts" table:
    select plan_id from accounts where exists (select * from parent_child, resellers where res_id = child_id and resellers.id = 'X' and account_id = accounts.id);
    where X is the reseller ID retrieved on step 3.
Compare the reseller ID retieved on step 5 with the one retrieved on step 4. If they don't coincide, update the "resellers" table:
update resellers set reseller_plan_id = 'X' where id = 'Y';
X is a reseller plan ID retrieved on step 5;
Y is a reseller ID retrieved on step 3.

15.4. When a user tries to create a new resource, it returns "Failed to add new resource over the hard credit".
This means Plan Credit Limit has been reached. User must have run out of money on his balance (if he pays by check), or his credit card failed to be charged (expired/was broken, etc.) In this case the user can still run his account (with recurent and extra usage payment acrued making up his negative balance), but creating new resources (both paid and free) is disabled.

15.5. We are having a double billing problem. If we manually charge an account, the system doubles it. Then if we credit it, it doubles the credit too.
Most likely two instances of your Admin Control Panel are running at the same time (either 2 instances of TomCat, but startup scripts were fixed in 2.08, or Tomcat and Apache Jserv). In the latter case, check your logs:
 grep FINISH /tmp/hsphere.log* 
If you see that two periods ended simultaniously, you are running two copies of your CP.
Alternatively, run the following command:
ps axuw | grep java
If you see both "DTomcat" and "../../HSdocumentation/FAQ/org.apache.JS" words in the line that appears, both are run by your system.
Next, stop the one you don't need or, which is more preferable, stop both and start either Apache or Tomcat.

15.6. User's credit limit doesn't match plan's default credit limit
To coincide with plan's default credit limit, individual account credit limit shoud be set to 0.

15.7. Credit card paid account has a negative balance but the user is not being charged.
Check if this negative balance stays within the credit limit. If the debt is over the credit limit, go to Reports -> Charge log and check for error messages from the processing center.

15.8. I receive trouble tickets with "Error: Cannot open new billing period".
Most probably the duration (size) of the billing period for the reported account is 0. To resolve the error, edit the billing period for this plan so its size is > than 0.
* If you don't want to charge this customer for hosting, switch his account to a plan without billing.

15.9. My customer has 1 database free. He purchased one more database, then deleted the first one, and got no refund.
This is the way H-Sphere billing is designed. The customer deleted a free resource. To avoid this problem, the customer had to delete the first database first and then create a new one. This is true of all H-Sphere resources.

16. Databases

16.1. phpMyAdmin is listing any database containing the word "order" instead of constraining the results to an EXACT match.
This is a known phpMyAdmin bug. A user can only see these databases, but can not do anythig with them.

16.2. Why neither dbs nor users are created under my end-user accounts?
If you have two *SQL logical servers available for signup, then the servers for dbs/logins/users creation are chosen by randomize. Imagine wwwuser creates database A and A1 and A2 users and they all go to sql1.mydomain.com. Then even if you prohibit it for signup, the system will still try to create new dbs and db users of wwwuser on sql1.mydomain.com. And as the signup is turned off on it, neither dbs nor users for this db will be created.

17. Trouble Tickets

17.1. Nobody of our support receives trouble tickets.
Make sure you have at least one account under the Tech Support Admin plan. It is designed specifically to receive and assign trouble tickets.

17.2. What should I do if I constanly get the following trouble ticket: @@@@@@@...WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED!
That might mean that:
a. OpenSSH keys were changed.
b. OpenSSH was reinstalled.
c. nethernet card was reinstalled/changed.
d. network interface was changed.

To fix this problem, please do the following:
1. log into the CP server under the cpanel user.
2. remove all the strings which contain the IP indicated in the message (as one that you failed to connect to with SSH) from the following two files:
a) ~cpanel/.ssh/known_hosts
b) ~cpanel/.ssh/known_hosts2

3. execute the following comand:
# ssh -l root IP (the one that RSA host key was changed for)
and agree to "Are you sure you want to continue connecting (yes/no)?"
Then make sure that the cpanel user can log into the server without any passwords and prompts.

17.3. I get 'empty' trouble tickets.
When your end user receives a response to their support request and responds to it not from within H-Sphere trouble ticket center, but through his e-mail client using the Reply option, make sure they place their message before, i.e. not below the ======== CUT HERE ========= quote, otherwise it won't show up in the body of the trouble ticket message in your trouble ticket center.

18. New Version Upgrades

18.1. You said you upgraded my H-Sphere. I cannot find new features in my control panel.
Possibly, you have created your own custom templates, and they are overriding the changes.

19. Customization

19.1. I can't switch between skins.
you need to enable skins as suggested here.

19.2. The custom templates don't take effect.
First, check if the file shiva/psoft_config/hsphere.properties is owned by cpanel. If it is, open this file and the USER_TEMPLATE_PATH line. Here, enter the full name of the directory with your custom templates, e.g.:
USER_TEMPLATE_PATH = /hsphere/local/home/shiva/custom/templates/.
Important: The directory name must end with a slash!
Don't do anything if the directory name is already there.
Then restart H-Sphere.

20. Suspending and Resuming Accounts

20.1. A reseller has suspended his own reseller admin account accidentally. How do I unsuspend this account?
1) Find this account by record account_id
2) log into posgres database
a. retrieve the password to the DB with:
cat ~cpanel/shiva/psoft_config/hsphere.properties| grep DB_PASSWORD
b. then execute:
psql -U wwwuser hsphere
c. enter the retrieved password
3) execute SQL command:
update accounts set suspended = null where id = X;
where X is the ID of the account that you want to resume; that will resume it.
4) execute: \q
to log off.
5) Restart H-Sphere.

20.2. I have suspended my master admin account. How do I resume it?
Log into your hsphere database and run the command:
update accounts set suspended=null where id=1;
Then restart H-Sphere.

20.3. Suspended user changed billing profile, but the account is still suspended.
H-Sphere doesn't resume accounts automatically. The customer should notify the admin that the profile has been changed,so the admin can resume the account manually.

21. Bandwidth or Traffic

21.1. Upon 'downgrading' plan, account traffic limit remained the same.
The user didn't agree to decrease his traffic limit to target plan's free amount when questioned while swithcing to the plan that provides less free traffic.

21.2. Upon changing account's plan the user is charged "Monthly fee for extra N Gb above free Y Gb of traffic".
The target plan provides less free traffic GBs than account's traffic limit. When changing plans user refused to automatically decrease his traffic limit to plans free as suggested in the warning message and now he either has to pre-pay for the allocation difference each traffic month or manually decrease his traffic limit to plan's free amount.

21.3. Although the user didn't exceed his month traffic limit, when changing it, he was accrued Usage fee for extra X GB above pre-paid Y GB of traffic.
From the start of the traffic month to the time when month traffic limit is changed the user ran up more (X GB) than traffic limit prorated to this period (Y GB).

21.4. Traffic month has passed, but Summary Traffic hasn't reset.
The user signed up with more than 1 month billing period later reconfigured as 1 month from admin control panel. That didn't change the end billing period date for the customer account and it stayed the same. Summary Traffic is a monthly resource and for accounts on 1 month billing period isn't accounted separately from the account usage charges at the end of billing period. The solution would be to change the account start billing date to the current date when you are changing it. That will close the current billing period, refund all the recurrent fees for the interrupted billing period the user signed up on and update the dates for a new one month billing period.

21.5. Mail/http/ftp traffic isn't calculated
Check if you have appropriate crons:
Login as root and execute:
crontab - l
it should list:
/hsphere/shared/scripts/cron/mail_anlz.pl on mail box
/hsphere/shared/scripts/cron/ftp_anlz.pl on web box
/hsphere/shared/scripts/cron/cron_analyze.pl on web box
'java psoft.hsphere.TrafficLoader' on cp box

21.6. Users are charged "Monthly fee for extra 0GB above free X GB of traffic".
If Traffic In use on the Online Price page in customer's account is bigger than their traffic limit, most likely there is more than one traffic resource in their account which causes extra charges. This can be fixed by re-setting customer's traffic limit to the same value.
* Improper recurrent payments can be credited back.

22. Urchin

22.1. Urchin returns the error "404 Not Found" when launching statistics calculation.
This means that the CGI script that prints out statistics (cgi-bin/print-log.pl for Unix and urchin\urchin.exe for Windows) is not found.
Please check your Web server configuration (cgi-bin configuration for Apache and urchin configuration for IIS). Also check if these CGI scripts are present in cgi-bin directories on the Web boxes where statistics is being calculated.
Log file with Urchin history: /hsphere/local/urchin/data/history

22.2. Urchin returns the error "500 Internal Error" when launching statistics calculation
If this happens under Unix, most probably, you have wrong permissions on the print-log.pl statistics collection script or on the loglist file.
print-log.pl should be owned by root:root and have 755 permissions.
loglist should be owned by httpd:httpd and have 644 permissions.
If this error occured under Windows, please contact support. We haven't faced this problem on Windows so far.

22.3. The statistics collection script has returned some data but not log files.
Most probably, an error had occured and the script returned the error message. According to the message, you may find the cause of the problem.


[16:44:05] Transferring;&3h3pydx11hvbqlod47crtw==&.gz...done (27 bytes).

[16:44:06] Logfile: /hsphere/local/urchin/var/logs/urchin-1092-0.log.gz
data lines: 0 (0%) data hits: 0
data proc: 0 bytes in 0 sec
data range: 0 - 0

Check the log file specified by the script:
cat /hsphere/local/urchin/var/logs/urchin-1092-0.log.gz
File loglist doesn't exists

This means that the loglist file hasn't been created. Create it in the same directory with print-log.pl. Set its owner as httpd:httpd and grant it 644 permissions:

cd /hsphere/local/sqwebmail/cgi-bin
touch loglist
chown httpd:httpd loglist
chmod 644 loglist

22.4. I have changed IP address of a server where Urchin is installed. How can I make changes to Urchin configuration?
You need to run the Urchin reconfiguration tool under cpanel user:
1) Log in as cpanel, e.g., under root:
su - cpanel
2) Run the tool that will make all necessary changes in Urchin profiles:
java psoft.hsphere.tools.UrchinReconfig

22.5. I have a problem with collecting log files. print-log.pl on a Unix Web server returns the "Wrong IP" error.
In this case, if there is no possibility to change network settings, create the urchin_ip_pool file in the same directory with print-log.pl. This file should have root:root permissions. There you put the list of IP addresses, each in a separate line, where Urchin would collect log files.

23. Windows

23.1. When enabling/disabling resources on Winbox, admin gets the following internal error:
Unable to create web site COM error: Failed to run metabase service API error : There are no more files.
An Error like this can be caused by a problem with IIS. If the Metabase has a wrong record, the Metabase service can't start. To fix this, remove this record and restart H-Sphere on WinBox.
To find the wrong record in the Metabase, restart H-Sphere and check the 'mbmgract.log' file (it is located in the '\HSphere\logs\' directory). It may contain the following records in the end of the file after you restarted H-Sphere:
CMetaBaseMgr::InitWebLists() - Ok
CMetaBaseMgr::InitFtpLists() - Ok
It means that H-Sphere has initialized Metabase. Otherwise, you get the number of a wrong record in Metabase, and you can remove it in the 'W3Svc' section using MetaEdit. Then you should restart IIS and H-Sphere and check the 'mbmgract.log' file again.

See also

Product Page
Doc Home
Online Demo
Installation Request
User's FAQ
Admin FAQ

Home   Products   Services   Partners   Support   News   Contact   Forum
© 2020 psoft.net.
All rights reserved.